A.I. UPDATE: Boom or Bubble?! | Lessons from Japan | Ep. 34

Welcome back from the future—literally. Three weeks in Japan revealed why the world's most advanced 1990s civilization got stuck in time while AI quietly achieved what seemed impossible just months ago.

This week, we examine what Japan teaches us about innovation stagnation, decode whether we're in an AI bubble or just moving goalposts, and reveal the medical breakthroughs happening behind the scenes while everyone debates model performance.

Let’s keep riding this tech wave together!

And don’t forget to follow us on socials and share share share all the innovation goodies!

Cheers,
Patrick, Nikola & Aragorn 🚀

Book Launch: Golven van Overvloed (Waves of Abundance)

On September 5th, together with Bot Uitgevers in Leiden, we proudly present the launch of “Golven van Overvloed” about Living and Working in a World of Exponential Growth. A story about the choice we all face: do we ride the waves of change, or risk being swept away by them?

We live in a time where standing still no longer offers safety, but rather a guarantee of going under. AI, robotics, virtual worlds, systems larger than ourselves, and even new forms of humanity, these are the five waves of change explored in the book. They affect our lives more deeply than most of us realize.

But this book is not about technology alone. It’s about people. About knowing yourself, developing yourself, and connecting with others in a world that is changing at an exponential pace. It’s about the choice to help shape the future, instead of being swept away by it.

This presentation is a gathering of curious minds—people who dare to ask questions, seek answers, and believe that together we can build something meaningful. Within the historic walls of Het Gravensteen—where centuries-old wisdom meets new insights—we will not only share stories but also create new connections.

Every guest will receive one of the very first signed copies of Waves of Abundance. More than that, you’ll meet other thinkers, builders, and dreamers who, just like you, are trying to understand where we are headed and how best to get there.

What to expect

The evening begins at 17:00 with a short presentation, followed by time to network and exchange ideas. Each guest will leave with a signed copy of the book. That’s also the reason for the ticket price. It simply covers your book! The waves of change won’t wait. The question is not if they are coming, but how you will choose to surf them. We look forward to welcoming you on September 5th in Leiden.

Reflections From Japan

Three weeks in Japan revealed a civilization frozen between technological excellence and innovation paralysis. "You feel like you're in a country where somehow they were super advanced technologically way ahead of rest of the world, but then suddenly it stopped," Aragorn observes about experiencing an alternate timeline where cash handling reached perfection, just as the world moved beyond cash.

Japan dominated global technology in the 1990s but hasn't meaningfully advanced since. The causes converge: economic collapse forced defensive thinking, aging CEOs refused to step aside for new generations, and hierarchical culture prioritized conformity over creativity. "They got stuck... They had economic collapse, which made them change from offense to defense."

The mindfulness and respect Patrick and Aragorn witnessed—no trash bins yet spotless streets, perfect public restrooms with linen towels, subway riders keeping phones silent—reflects deep cultural values. But the same system thinking that creates social harmony also prevents the flexibility required for innovation.

Why does it matter? 
Japan demonstrates how success can become a trap. Organizations that prioritize perfecting existing systems over adapting to new realities risk technological obsolescence, no matter how advanced their current capabilities. The lesson for the AI era: maintaining offensive innovation mindset becomes crucial as capabilities rapidly evolve.

The AI Bubble Myth

Critics increasingly claim we're in an "AI bubble" similar to the dot-com crash, but the numbers tell a different story. Current estimates suggest 1.7-1.8 billion people globally have used AI tools, with 500-600 million engaging daily and ChatGPT alone serving 700 million weekly active users.

But this definition focuses purely on financial metrics, not technological success. The dot-com "bubble" delivered the internet revolution exactly as predicted—the only failure was investor timing expectations.

“Five years ago, people believed that the chances of us getting to AGI within our lifetime was less than 1%. Now, we actually have an AI bubble because GPT-5 is not artificial general intelligence yet. Are we crazy?". In a nutshell, expectations shifted from "impossible in our lifetime" to "disappointed it's not here yet."

The 1.8 billion user adoption rate represents unprecedented technology uptake—faster than the internet, faster than mobile phones, faster than any previous innovation in human history. For comparison, LinkedIn took 15 years to reach 800 million users; AI achieved similar numbers in under two years.

Why does it matter?
The "bubble" exists primarily in investor psychology and moving expectations, not technological reality. While financial corrections may occur, the underlying AI progress continues independently of stock market sentiment. Organizations focused on practical AI implementation rather than speculation will navigate any financial turbulence successfully.

GPT-5's Anticlimax

OpenAI's GPT-5 release generated more disappointment than excitement, scoring marginally better than competitors on benchmarks while feeling incrementally rather than revolutionarily improved.

Major AI models now perform within 1% of each other across most tasks. The competitive advantage has shifted from raw capability to accessibility, features, and user experience. The underwhelming release could indicate either a technical plateau requiring new breakthroughs, or strategic restraint. Political considerations under the Trump administration may influence release decisions for national security reasons.

Expected multimodal improvements didn't materialize, possibly due to deteriorating Microsoft-OpenAI relations limiting system integrations. Meanwhile, Google continues gradually expanding Gemini's capabilities with email search, image editing, and deeper system integration.

Why does it matter?
The model convergence signals a maturation phase where implementation and integration matter more than benchmark performance. Organizations should focus on workflow optimization and practical applications rather than chasing the latest model releases. The real competition now occurs in ecosystem integration, not isolated model capabilities.

AlphaFold 4: The Protein Revolution Accelerates

Google DeepMind released AlphaFold 4 in July, achieving molecular-scale accuracy in protein structure prediction that enables atomic-level experimental research. The breakthrough represents the fourth major iteration of technology that has fundamentally transformed biological research in just four years.

AlphaFold 4 can predict protein structures with "pinpoint accuracy" at the molecular level, making it completely viable for experimental research and atomic-level reconstruction by scientists. The technology is now open-sourced, enabling laboratories worldwide to access capabilities that previously required massive institutional resources.

The gap between breakthrough and application remains significant due to regulatory approval processes unchanged by AI. "It can take two to five years before these medications get rolled out. But that does mean that in two to five years, we might suddenly have a wave of new medication that cures cancer, basically stops you from aging."

Why does it matter?
AlphaFold represents AI's most concrete impact on human welfare—not through productivity improvements but through fundamental scientific breakthroughs that will extend and improve human life. Even if financial AI bubbles burst, these discoveries continue developing independently, ensuring transformative medical advances regardless of market sentiment.

Reversing Aging: AI Redesigns Life Itself

Scientists used specialized AI to reverse aging in human cells, achieving 50x better efficiency than the original Nobel Prize-winning method.

OpenAI partnered with longevity startup RetroBio to redesign Yamanaka factors—proteins that can reprogram adult cells back to youthful stem cells.

Traditional methods converted less than 0.1% of cells over weeks. The AI-redesigned proteins achieved over 30% success rates with the best version showing 50-fold increases in stem cell markers and dramatically enhanced DNA damage repair.

GPT-4B-Micro was trained exclusively on biological data, enabling it to redesign entire proteins with hundreds of simultaneous changes rather than testing individual mutations. This specialized approach compressed decades of potential research into weeks.

Full breakdown on OpenAI’s first biological AI model 👇️ 

The success demonstrates how domain-specific AI training can achieve breakthrough results in specialized fields. Rather than general-purpose models, the future may belong to highly specialized AI systems trained on narrow but deep datasets.

Why does it matter?
This represents AI's transition from information processing to biological engineering. The technology doesn't just analyze existing knowledge—it creates new solutions that surpass human-designed alternatives. Organizations should prepare for AI that doesn't just automate existing processes but fundamentally redesigns them for superior performance.

Open CRISPR-1: Programming Life Like Code

Open CRISPR-1 enables AI-assisted DNA programming, allowing researchers to "vibe code DNA" the same way developers write software. The technology combines CRISPR gene editing capabilities with large language models to create entirely new biological systems rather than just modifying existing ones.

Unlike traditional CRISPR that makes precise cuts to existing genes, Open CRISPR-1 enables designing completely novel genetic sequences.

The technology requires only internet access and basic laboratory equipment costing thousands rather than millions.

"One crazy person that uses the AI to become superhuman in some way, to do something that was previously only capable or accessible to multinationals or nations that can now be done by a crazy individual in shed in the back of their garden."

Why dost it matter?
Open CRISPR-1 represents a fundamental shift in biological capabilities from institutional to individual access. Unlike atomic weapons requiring rare materials and massive infrastructure, genetic engineering now requires only knowledge and basic equipment. This democratization necessitates new governance frameworks for biotechnology as powerful as nuclear energy but accessible as software development.

Honorable Mentions

AI is decoding previously mysterious "junk DNA" regions, discovering critical epigenetic switches and regulatory functions that change throughout human lifetimes, revolutionizing our understanding of genetic expression.

Google's Genie 3 world-generation AI creates explorable environments in real-time, including recursive worlds within worlds visible through mirrors and windows—essentially AI-powered Inception scenarios that preview the metaverse's future.

Google is mapping and summarizing every 10-square-kilometer area of Earth, creating an AI model representing the entire planet that you can query conversationally about any location or environmental condition.

The first fully autonomous surgical operation was performed successfully on a pig cadaver, representing a major step toward AI surgeons that could replace human doctors within a decade.

New AI stethoscopes can detect heart anomalies in 15 seconds that human doctors cannot perceive, enabling earlier intervention and better outcomes for cardiovascular conditions.

Newsletter full video recording